A library of biorthogonal wavelet transforms originated from polynomial splines

Size: px
Start display at page:

Download "A library of biorthogonal wavelet transforms originated from polynomial splines"

Transcription

1 A library of biortogonal wavelet transforms originated from polynomial splines Amir Z. Averbuc a and Valery A. Zeludev a a Scool of Computer Science, Tel Aviv University Tel Aviv 69978, Israel ABSTRACT We present a library of biortogonal wavelet transforms and te related library of biortogonal symmetric waveforms. For te construction we use interpolatory, quasiinterpolatory and smooting splines wit finite masks (local splines). On tis base we designed a set of perfect reconstruction infinite and finite impulse response filter banks wit te linear pase property. Te construction is performed in a lifting manner. Te developed tecnique allows to construct wavelet transforms wit arbitrary prescribed properties suc as number of vanising moments, sape of wavelets, frequency resolution. Moreover, te transforms contain some scalar control parameters wic enable teir flexible tuning in eiter time or frequency domains. Te transforms are implemented in a fast manner. Te transforms based on interpolatory splines are implemented troug recursive filtering.we present encouraging results of applicarion of devised wavelet transforms to image compression.. BIORTHOGONAL WAVELET TRANSFORM Te sequences {a(k)} k=, wic belong to te space l, we will call te discrete-time signals. discrete-time signals we denote by S. Te z transform of a signal {a(k)} S is defined as follows: Te space of a(z) = k= z k a(k). Trougout te paper we assume tat z = e iω. We introduce a family of biortogonal wavelet-type transforms tat operate on te signal x = {x(k)} k=, wic we construct troug lifting steps. First we recall a basic lifting sceme of wavelet transform. Te lifting sceme can be implemented in a primal or dual modes. For brevity we consider only te primal mode... Lifting sceme of wavelet transform.. Decomposition Generally, te primal lifting sceme for decomposition of signals consists of tree steps:. Split.. Predict. 3. Update or lifting. Split - We split te array x into an even and odd sub-arrays: e = {e (k) = x(k)}, d = {d (k) = x(k + )}, k Z. Predict - We use te even array e to predict te odd array d and redefine te array d as te difference between te existing array and te predicted one. To be specific, we apply some filter wit transfer function zu(z) to te sequence e and predict te function d (z ) wic is te z transform of d. Te z transform of te new d array is defined as follows: d u (z ) = d (z ) zu(z)e (z ). (.) From now on te superscript u means an update operation of te array. Obviously, zu(z)e (z ) sould well approximate d (z ). Furter autor information: (Send correspondence to Amir Z. Averbuc ) Amir Z. Averbuc: amir@mat.tau.ac.il Valery A. Zeludev : zel@mat.tau.ac.il

2 Lifting - We update te even array using te new odd array: e u (z ) = e (z ) + β(z)z d u (z ). (.) Generally, te goal of tis step is to eliminate aliasing wic appears wile downsampling te original signal x into e. By doing so we ave tat e is transformed into a low-pass filtered and downsampled replica of x. Furter on we will discuss ow to acieve tis effect by a proper coice of te control filter β, but for now we only require te function β(z) to be real-valued and obey te condition β( z) = β(z), so te product β(z)z is a function of z.... Reconstruction Te reconstruction of te signal x from te arrays e u and d u is implemented in reverse order:. Undo Lifting.. Undo Predict. 3. Unsplit. Undo Lifting - We restore te even array: e (z ) = e u (z ) β(z)z d u (z ). Undo Predict - We restore te odd array: d (z ) = d u (z ) + zu(z)e (z ). Unsplit - Te last step represents te standard restoration of te signal from its even and odd components. In te z domain it looks as: x(z) = e (z ) + z d (z )..3. Filter banks Lifting scemes, tat were presented above, yield efficient algoritms for te implementation of te forward and backward transform of x e u d u. But tese operations can be interpreted as transformations of te signals by a filter bank tat possesses te perfect reconstruction properties. First we define two filter transfer functions Φ l (z) = ( + U(z))/, Define te filter functions g(z) = z Φ (z), Φ (z) = ( U(z))/. β (z) = + β(z)φ (z) = + β(z) z g(z), (z) = Φ l (z) g β (z) = z ( β(z)φ l (z) ) = z ( β(z) (z)). We call β (z) and g(z) te transfer functions of te low-pass and ig-pass primal decomposition filters, respectively. We call (z) and g β (j) te transfer functions of te low-pass and ig-pass primal reconstruction filters, respectively. Tese four filters form a perfect reconstruction filter bank (Ref. 5). Teorem.. Te decomposition and reconstruction formulas can be represented as follows: e u (z ) = (z) x(z) + ( β ) β ( z) x( z) d u (z ) = ( ) g(z) x(z) + g r ( z) x( z) (.3) x(z) = (z)e u (z ) + g β (z) d u (z ). (.) If β( z) = β(z) ten te functions β (j), g(j), (j) and g β (j) satisfy te perfect reconstruction conditions (z) β (z) + g β (z) g(z) = (z) β ( z) + g β (z) g( z) =. Te filter functions are linked to eac oter in te following way: g(z) = z ( z); g β (z) = z β ( z)

3 .. Bases for te signal space Te perfect reconstruction filter banks, tat were constructed above, are associated wit te biortogonal pairs of bases in te space S of discrete-time signals. In section.3 we introduced a family of filters by teir transfer functions (z), g β (z), β (z), g(z). We denote by ϕ (k), ψ β (k), ϕ β (k), ψ (k) te impulse response functions of te corresponding filters, respectively. It means tat, for example r (z) = k Z z k ϕ r, (k) and similarly for te oter functions. Proposition.. Te sifts of te functions ϕ (k), ψ β (k), ϕ β (k) and ψ (k) form a biortogonal pair of bases for te signal space S. Tis means tat any signal x S can be represented as: x(l) = e u (k) ϕ (l k) + d u ψβ(l k). k Z k Z Te coordinates e u (k) and d u (k) can be represented as inner products: e u (k) = x, ϕ β,k, were ϕ β,k(l) = ϕ β(l k) d u (k) = x, ψ k, were ψ k(l) = ψ (l k). Te following biortogonal relations old: ϕ β,k, ϕ l = ψ β,k, ψ l = δ l k, ϕ β,k, ψ β,l = ψ l, ϕ k =, l, k. Te formulated proposition justifies te following definition. Definition.. Te functions ϕ and ψβ, wic belong to te space S, are called te low-frequency and igfrequency syntesis wavelets of te first scale, respectively. Te functions ϕ β and ψ are called te primal lowfrequency and ig-frequency analysis wavelets of te first scale, respectively. Repeated applications of te transform can be acieved in an iterative way. In tis transform we store te array d u and decompose te array e u. Te transformed arrays e u and d u of te second decomposition scale are derived from te even and odd sub-arrays of te array e u by te same lifting steps as tose described above.i As a result we get tat te signal x is transformed into tree subarrays: x d u d u e u. Te reconstruction is performed in te reverse order.. POLYNOMIAL SPLINES We will construct polynomial splines of various kinds using te even subarray of a signal, calculate teir values in te midpoints between nodes and use tese values for prediction of te odd array. In tis section we discuss some properties of suc splines and derive te corresponding filters U... B-splines Te central B spline of te first order on te grid {k} is defined as follows: { M / if x [ /, /] (x) = elsewere. Its Fourier transform is: M (ω) = e iωx M (x)dx = V (ω), V (ω) = sin(ω/) ω/ (.5) Te central B spline of order p is te convolution M p p (x) = M (x) M(x) = π e iωx (V (ω) p dω p. (.6)

4 Note tat te B-spline of order p is supported at te interval ( p/, p/). It is positive witin its support and symmetric around zero. Te nodes of B-splines of even orders are located at points {k} and of te odd orders at points {(k + /)}, k Z. It follows from (.6) tat M p (x) = M p (x), be presented explicitly. Suppose x + = (x + x )/. Ten M p p ( ) p ( (x) = ( ) k x k + p (m )! k We introduce te following sequences k= u p = {M p (k) = M p (k)}, w p = {M p were M p (x) = M p (x). B-splines can ) p +. (.7) ((k + ) ( ) = M p k + )}, k Z (.8) comprised of values of B-splines in te grid points and in te midpoints. Due to te compact support of B-splines, tese sequences are finite. By te reasons to be explained furter on, we will use for our constructions only splines of odd orders p = r. In Tables, we present te sequences for initial values r wic are of practical importance. We need te z transforms of te sequences u p and w p : k u u u Table. Values of te sequences u p. k w 3 w 5 w Table. Values of te sequences w p. u p (z) = k= z k u p (k), w p (z) = k= z k w p (k). Tese functions are Laurent polynomials. Tey are called te Euler-Frobenius polynomials Ref. and were extensively studied in Ref. 3. In particular, te following important facts Ref. 3 were discovered: Proposition.. On te circle z = e iω te Laurent polynomials u p (z) are strictly positive. Teir roots are all simple and negative. Eac root ζ can be paired wit a dual root θ suc tat ζ θ =. Tus if p = r + is odd ten u p (z) can be represented as follows: were < γ n <. We denote u p (z) = r n= γ n ( + γ n z )( + γ n z ), (.9) U p i (z) = z wp (z) u p (z). (.) Proposition.. Te rational functions U p i (z) are real-valued and U p i ( z) = U p i (z). If p = r + is odd ten U p i (z) = (α )r+ ξ r (α) u p (z) were α = z + z and ξ r (α) is a polynomial of degree r., + U p i (z) = ( α )r+ ξ r ( α) u p, (.) (z)

5 .. Interpolatory splines Te sifts of B-splines form a basis in te space S p of splines of order p on te grid k. Namely, any spline Sp Sp as te following representation: S p (x) = q(l) M p (x l). (.) l Denote q = {q(l)} and let q(z ) be te z -transform of q. We introduce also te sequences s p = {S p (k) = Sp (k)} and m p = {S p ((k + /)) = Sp (k + /)} of values of te spline in te grid points and in te midpoints. Let s p (z ) and m p (z ) be corresponding z -transforms. We ave S p (k) = l q(l) M p (k l), Sp ( k + ) = l q(l) M p ( k l + ). (.3) Respectively, s p (z ) = q(z )u(z), m p (z ) = q(z )w(z). From tese formulas we can derive expression for te coefficients of a spline wic interpolates a given sequence e = {e(k)} at grid points: Te z transform of te sequence m p is: S p (k) = e(k), k Z, q(z )u p (z) = e(z ) q(z ) = e(z ) u p (z). (.) m p (z ) = q(z )w p (z) = zu p i (z)e(z ). (.5) Our furter construction exploits te super-convergence property of te interpolatory splines of odd orders (even degrees). Teorem. (Ref. 7). Let a function f L (, ) as p + continuous derivatives and S p Sp interpolates f on te grid {k}. Denote f k = f(k +/)). Ten in te case of odd p = r +, te following asymptotic relation olds S p ((k + /) = f k r+ f (r+) ((k + /)(r + ) b r+() b r+ (/) (r + )! were b s (x) is te Bernoulli polynomial of degree s. + o( r+ f (r+) (x)), (.6) Recall, tat in general te interpolatory spline of order r + approximates te function f wit accuracy of r+. Terefore, we may claim tat {(k + /)} are points of super-convergence of te spline S p. Note, tat te spline of order r + wic interpolates te values of a polynomial of degree r coincides wit tis polynomial. However, te spline of order r + wic interpolates te values of a polynomial of degree r + on te grid {k} restores te values of tis polynomial in te mid-points {(k + /)}. Tis property will result in te vanising moments property of te wavelets to be constructed furter on..3. Quasi-interpolatory splines It is seen from Eqs. (.), (.5) tat, in order to find coefficients of te spline interpolating a signal e and te values of te spline in te midpoints, te signal as to be filtered wit te filters, wose transfer functions are /u p (z) and zu p i (z), respectively. Tese filters ave infinite impulse responses (IIR). However, te property of super-convergence in te midpoints is not exclusive attribute of te interpolatory splines. It is also inerent to te so called local quasi-interpolatory splines of odd orders, wic can be constructed using finite impulse response (FIR) filtering. Definition.. Let a function f as p continuous derivatives and f = {f k = f(k)}, k Z. A spline S p Sp of order p given by (.) is said to be te local quasi-interpolatory spline if te array q of its coefficients is derived by FIR filtering te array of samples f q(z ) = Γ(z )f(z ), (.7)

6 were Γ(z ) is a Laurent polynomial, and te difference f(x) S p (x) = O(f (p) p ).Provided f is a polynomial of degree p, te spline S p (x) f(x). If w p is te sequence defined in (.8) ten te values of te spline S p in te midpoints between grid points mp are produced by te following FIR filtering te array of samples f m p (z ) = zu p (z)f(z ), U p (z) = z Γ(z )w p (z ). (.8) In Ref. 7 explicit formulas for construction of quasi-interpolatory splines are establised as well as te estimations of te differences. In te present work we are interested in te splines of odd orders p = r +. Tere are many FIR filters wic generate quasi-interpolatory splines but only one filter of minimal lengt r + for eac order p = r +. Denote λ(z) = z + z.. Teorem.3. A quasi-interpolatory spline of order p = r + can be produced by filtering (.7) wit filters Γ of lengt no less tan r +. Tere exists a unique filter Γ r m of lengt r +, wic produces te minimal quasi-interpolatory spline (x). Its transfer function is: S r+ Γ r m(z ) = + r βkλ r k (z). (.9) k= Te coefficients β r k can be derived from te generating function ( ) r+ arcsin t/ = ( ) k β r t kt k. (.) If te function f as r + 3 derivatives ten te following asymptotic relations old for te minimal quasiinterpolatory spline S r+ (x) for x = (k + / + τ), τ [, ]: S r+ (x) = f(x) r+ f (r+) (x) b ( ) r+(τ) (r + (r + )! + r+ f (r+) )br+ (τ) (x) βr+ r + O(f (r+3) r+3 ), (r + )! (.) were b s (x) is te Bernoulli polynomial of degree s. We recall tat te values b s = bs () are called te Bernoulli numbers. If s is odd and s > ten b s =. Hence, values of te minimal quasi-interpolatory spline in te midpoints between nodes are te following : S r+ ((k + )) = f((k + )) + r+ f (r+) ((k + ))Ar + O(f (r+3) r+3 ), A r (r + )b r+ = β r (r + )! r+.(.) Tis implies te super-convergence property similar to tat of te interpolatory splines. Te minimal quasiinterpolatory spline of order r + (of degree r +) restores in te midpoints between nodes polynomials of degree r +. Te asymptotic representation (.) allows to enance te approximation accuracy of te minimal spline in te midpoints upgrading te filter Γ r m. Proposition.3. If te coefficients of a spline Šr+ S r+ of order r + are derived in line wit (.7) using te filter Γ r e of lengt r + 3 wit te transfer function Γ r e(z ) = Γ r m(z ) A r λ r+ (z) ten te spline restores polynomials of degree r + 3 in te midpoints between nodes. Tere exists one more way to upgrade te filter Γ r m, wic retains te approximation accuracy of te minimal spline in te midpoints but allows custom design of splines. Proposition. ( Ref. 8). If te coefficients of a spline S r+,ρ S r+ of order r + are derived in line wit (.7) using te filter Γ r ρ of lengt r + 3 wit te transfer function Γ r ρ(z ) = Γ r m(z ) + ρλ r+ (z) ten te spline restores polynomials of degree r + in te midpoints between nodes for any real value ρ. If te parameter ρ is cosen suc tat ρ = ( ) r ρ tan te spline S r+,ρ possesses te smooting property. It means tat it approximates values of te function f in te midpoints between nodes being constructed from te noised samples of te function: f = { f k = f(k) + ε k }, k Z. k=

7 .. Examples First we present a few values of te coefficients β r k : β r = r +, βr = (r + )(r + 7) 576, β3 r = (r + )(r + 6r + 5). 93 For te construction of te extended splines te values of te Bernoulli numbers b s are required. Te initial values are: b = /6, b = /3, b 6 = /.... Quadratic splines Interpolatory spline Denote α = z + z. U i (z) = α z z, U i (α ) (z) = z z, (.3) Minimal spline Te filters are. Γ m(z ) = 8 λ(z), U m(z) = z 3 + 9z + 9z z 3 6, Um(z) = (α ) (z + + z), (.) 6 Extended spline Γ e(z) = Γ m(z ) λ (z), Ue (z) = 3z 5 5z 3 + 5z + 5z 5z 3 + 3z 5, (.5) 56 Ue (z) = (α )3 (3z + 8z z + 3z ), 56 Remark Donoo Ref. presented a sceme were an odd sample is predicted by te value in te central point of te polynomial of an odd degree wic interpolates adjacent even samples. One can observe tat our filter Um coincides wit te filter derived by Donoo s sceme using te cubic interpolatory polynomial. Te filter Ue coincides wit te filter derived using te interpolatory polynomial of fift degree. On te oter and te filter Ui is closely related to te commonly used Butterwort filter Ref.. Namely, in tis case te filter transfer functions Φ,l i (z) = ( + Ui (z))/, Φ, i (z) = ( Ui (z))/. coincide wit magnitude squared te transfer functions of te discrete-time low-pass and ig-pass alf-band Butterwort filters of order, respectively.... Splines of fift order (fourt degree) Interpolatory spline Ui (z) = 6(z3 + z + z + z 3 ) z + 76z z + z, U i (α ) 3 (α ) (z) = z + 76z z. (.6) + z Minimal spline Te filters are Γ m(z ) = 5 7 λ(z) + 5 λ (z), Um(z) = 7(z 7 + z 7 ) + 89(z 5 + z 5 ) 77(z 3 + z 3 ) α 768 Um(z) = (α )3 (7(z + z ) + 8(z 3 + z 3 ) + 76(z + z ) + 366α

8 3. WAVELET TRANSFORMS USING SPLINE FILTERS 3.. Coosing te filters for te lifting step In te previous section we presented a family of te filters U for te predict step wic originate from splines of various types. But, as it is seen from (.), to accomplis te transform we ave to define te filter β. Tere is a remarkable freedom of te coice of tese filters. Te only requirement needed to guarantee te perfect reconstruction property of te transform is β( z) = β(z). In order to make syntesis and analysis filters similar in teir properties, we coose β(z) = Ŭ(z)/, were Ŭ means one of filters U presented above. In particular, Ŭ may coincide wit te filters U wic was used for te prediction. We say tat a wavelet ψ as m vanising moments if te following relations old: k Z ks ψ(k) =, s =,,..., m. Proposition 3.. Suppose tat for te predict and lifting steps filters U(z) and β(z) = Ŭ(z)/ are used, respectively. If U(z) comprises te factor (z + /z) r ten te ig-frequency analysis wavelets ψ ave r vanising moments. If, in addition Ŭ(z) comprises te factor (z + /z)p ten te syntesis wavelet ψβ as q vanising moments, were q = min{p, r}. 3.. Implementation of transforms Once we ave cosen te filter β = Ŭ/, te decomposition procedure is te following (see (.), (.)): d u (z ) = d (z ) zu(z)e (z ), e u (z ) = e (z ) + z Ŭ(z) d u (z ). (3.7) Te transfer function z Ŭ(z)/ differs from zŭ(z)/ only by te factor z, tat is te impulse responses of te corresponding filters are identical up to a sift. Tus bot operations of te decomposition are, in principle, identical. For te reconstruction te same operation are conducted in a reverse order. Terefore, it is sufficient to outline implementation of filtering wit te function zu(z). Tis function depends on z and we denote it as F (z ) = zu(z). Ten te decomposition formulas (3.7) are equivalent to te following d u (z) = d (z) F (z)e (z), e u (z) = e (z) + z F (z) d u (z). (3.8) Equation (3.8) means tat in order to obtain te detail array d u, we must process te even array e by te filter F wit te transfer function F (z) and extract te filtered array from te odd array d. In order to obtain te smooted array e u, we must process te detail array d u by te filter Φ tat as te transfer function Φ(z) = z F (z)/ and add te filtered array to te even array e. But te filter Φ differs from F r / only by one-sample delay and it operates similarly. Implementation of FIR filters originated from local splines is straigtforward and, terefore we only make a few remarks on IIR filters originated from interpolatory splines. Equations (.9) and (.) imply tat, wile te interpolatory spline of order r + is used, te transfer function P (z) F (z) = r n= γ n ( + γ n z)( + γ n z ), (3.9) were P (z) is a Laurent polynomial. It means tat te IIR filter F can be split into a cascade consisting of a FIR filter wit te transfer function P (z), r elementary causal recursive filters, denoted by R(n), and r elementary anti-causal recursive filters, denoted by R(n). On te oter and, te filters can be decomposed into sums of elementary recursive filters. Suc a decomposition also allows parallel implementation of te transform. Te causal and anti-causal filters operate as follows: y = R(n)x y(l) = x(l) + γ n y(l ), y = R(n)x y(l) = x(l) + γ n y(l + ). (3.3)

9 3.3. Examples of recursive filters Now we present IIR filers derived from te interpolatory splines of tird and fift orders. r =. Denote γ = 3.7. Ten Fi (z) = γ + z ( + γ z)( + γ z ). Te filter can be implemented wit te following cascade: x (k) = γ (x(k) + x(k + )), x (k) = x (k) γ x (k ), y(k) = x (k) γ y(k + ). Anoter option stems from te following decomposition of te function F i (z): ( Fi (z) = γ + γ Ten te filter is implemented in parallel mode: + γ + z z + γ z y (k) = x(k) γy (k ), y (k) = x(k + ) γy (k + ), y = γ + γ (y + y ). We note tat elementary filters, wic produce y and y, are operating in opposite directions. r =. Denote ξ = 5 6, γ = , γ = So, we ave te cascade representation: F i (z) = 6γ α ξ + ξz + γ z + ξz + γ z + γ z ). + γ ( + z). z. APPLICATION TO IMAGE COMPRESSION We carried out a series of experiments on compression of various images using a number of wavelet transforms constructed above... Employed wavelet transforms We explored various combinations of filters for te prediction and lifting steps obtaining different transforms. Most of tem are performing satisfactory on compression of images. We display in te following figures wavelets and teir spectra for te tree types of transforms: Ti Te transform were for te prediction and lifting steps te IIR filters U i are used (see (.3) We display in Figure te syntesis and analysis wavelets related to te transform Ti and teir Fourier spectra. One can observe tat te analysis wavelets are regular and, to some extent, are similar to te syntesis ones. Bot te analysis and syntesis ig-frequency wavelets ave vanising moments. Tm Te transform were for te prediction and lifting steps te FIR filters Um are used (see (.) We display in Figure te compactly supported syntesis and analysis wavelets related to te transform Tm and teir Fourier spectra. One can observe tat, unlike te analysis wavelets, te syntesis wavelets are regular and are similar to te syntesis wavelets sown in Figure ones. Bot te analysis and syntesis ig-frequency wavelets ave vanising moments. Ti Te transform were for te prediction and lifting steps te IIR filters U i are used (see (.6) We display in Figure 3 te syntesis and analysis wavelets related to te transform Ti and teir Fourier spectra. One can observe tat te analysis wavelets are regular and are similar to te syntesis ones. Bot te analysis and syntesis ig-frequency wavelets ave 6 vanising moments.

10 Figure. Te left pair bottom to top: Te syntesis wavelets used by te transform Ti : ϕ and ψβ k, k =, 3,, and teir Fourier spectra. Te rigt pair bottom to top: Te analysis wavelets used by te transformti : ϕ β and, k =, 3,, and teir Fourier spectra. ψ k β Figure. Te left pair bottom to top: Te syntesis wavelets used by te transform Tm: ϕ and ψβ k, k =, 3,, and teir Fourier spectra. Te rigt pair bottom to top: Te analysis wavelets used by te transformti : ϕ β and, k =, 3,, and teir Fourier spectra. ψ k β Figure 3. Te left pair bottom to top: Te syntesis wavelets used by te transform Ti : ϕ and ψβ k, k =, 3,, and teir Fourier spectra. Te rigt pair bottom to top: Te analysis wavelets used by te transformti : ϕ β and, k =, 3,, and teir Fourier spectra. ψ k β.. Experiments wit image compression We conducted a wide series of experiments wit compression of multimedia images using te above transforms. In most experiments tey outperform te popular B9/7 biortogonal filter bank. In tis section we present te results after compression and decompression of of four images. Two of tem are commonly used images Lena and Barbara and oter two Car and Fabrics are displayed in Figure. Tese are bit (8bpp) images. Te following experiments were done:

11 Figure. Left: Car, rigt: Fabrics.. We decomposed te image wit te wavelet transform up to 6-t scale using te B9/7 filters and te filters listed in Section... Te transform coefficients were coded using te SPIHT algoritm by Said and Pearlman Ref. 6. Tis algoritm enables to compress data exactly to a prescribed rate. We coded te coefficients wit compression ratios (CR) : (.8 bpp), : (. bpp), :3 (/5 bpp), : (. bpp) and :5 (/5 bpp). 3. Te reconstructed image was compared wit te original image and te peak signal to noise ratio (PSNR) in decibels was computed. ( ) N 55 P SNR = log N k= (x(k) db, x(k)) Lena and Barbara: Te PSNR values of Lena and Barbara images are presented in Table 3. One can observe Lena CR B9/7 Ti Tm Ti Barbara CR B9/7 Ti Tm Ti Table 3. PSNR of te Lena (left table) and Barbara (rigt table) images after te application of SPIHT were te decomposition of te wavelet transforms was into 6 scales. tat for Lena te transform Ti outperforms te B9/7 filter in any compression rate. Te computational complexity of tis transform is sligtly iger tan tat of te B9/7 filter but using parallel mode it can be implemented faster. Te FIR transform Tm, outperform te B9/7 filters in all CR except. Te computational complexity of tis transform is te same as tat of te B9/7 filter but te advantage is tat Tm can be implemented in integers. Computationally more expensive transform Ti in te parallel mode can be implemented faster tan te transform wit B9/7 filter. It is most efficient in te sense of PSNR. Even more it is true for te compression of Barbara. Here Ti remarkably outperforms all oter transforms. Te Tm transform gives way to B9/7. Car and Fabrics : Te PSNR values of Car and Fabrics are presented in Table. On bot images Ti transform outperforms te B9/7 filter for any compression ratio. On Car, wic comprise ig-frequency components,te FIR transform Tm proved to be efficient.

12 Car CR B9/7 Ti Tm Ti Fabrics CR B9/7 Ti Tm Ti Table. PSNR of te Car (left table) and Fabrics (rigt table) images. 5. CONCLUSIONS In tis paper we proposed an efficient metod tat produces new filters for wavelet transforms. Tese wavelet transforms were designed by te usage of polynomial interpolatory and quasi-interpolatory splines. Being applied to te compression of multimedia images, tese filters outperform te traditional biortogonal 9/7 filters wic are frequenty used in wavelet based compression. Te new filters and te biortogonal 9/7 are incorporated into SPIHT in order to measure and compare teir performance wit one well known codec. We presented also wavelet filters wit controlling parameter, wic can be used, in particular for processing noised signals and images. In te future we plan to explore tis opportunity and to ceck te performance of devised filters on video and seismic data. REFERENCES. D. L. Donoo, Interpolating wavelet transform, Preprint 8, Department of Statistics, Stanford University, 99.. I. J. Scoenberg, Cardinal spline interpolation, CBMS,, SIAM, Piladelpia, I. J. Scoenberg, Contribution to te problem of approximation of equidistant data by analytic functions, Quart. Appl. Mat., (96), 5-99, -.. A. V. Oppeneim, R. W. Safer, Discrete-time signal processing, Englewood Cliffs, New York, Prentice Hall, G. Strang, and T. Nguen, Wavelets and filter banks, Wellesley-Cambridge Press, A. Said and W. W. Pearlman, A new, fast and efficient image codec based on set partitioning in ierarcical trees, IEEE Trans. on Circ. and Syst. for Video Tec., 6: 3-5, June V. A. Zeludev Local spline approximation on a uniform grid, U.S.S.R. Comput. Mat. & Mat. Pys., 7 No.5, (987), V.A. Zeludev, Local smooting splines wit a regularizing parameter, Comput. Mat. & Mat Pys., 3, (99), 93-.

Wavelet transforms generated by splines

Wavelet transforms generated by splines Wavelet transforms generated by splines Amir Z. Averbuch Valery A. Zheludev School of Computer Science Tel Aviv University Tel Aviv 69978, Israel Abstract In this paper we design a new family of biorthogonal

More information

Linear Interpolating Splines

Linear Interpolating Splines Jim Lambers MAT 772 Fall Semester 2010-11 Lecture 17 Notes Tese notes correspond to Sections 112, 11, and 114 in te text Linear Interpolating Splines We ave seen tat ig-degree polynomial interpolation

More information

Computation of interpolatory splines via triadic subdivision

Computation of interpolatory splines via triadic subdivision Computation of interpolatory splines via triadic subdivision Valery A. Zheludev and Amir Z. Averbuch Abstract We present an algorithm for computation of interpolatory splines of arbitrary order at triadic

More information

PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION

PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION PYRAMID FILTERS BASED ON BILINEAR INTERPOLATION Martin Kraus Computer Grapics and Visualization Group, Tecnisce Universität Müncen, Germany krausma@in.tum.de Magnus Strengert Visualization and Interactive

More information

Piecewise Polynomial Interpolation, cont d

Piecewise Polynomial Interpolation, cont d Jim Lambers MAT 460/560 Fall Semester 2009-0 Lecture 2 Notes Tese notes correspond to Section 4 in te text Piecewise Polynomial Interpolation, cont d Constructing Cubic Splines, cont d Having determined

More information

Numerical Derivatives

Numerical Derivatives Lab 15 Numerical Derivatives Lab Objective: Understand and implement finite difference approximations of te derivative in single and multiple dimensions. Evaluate te accuracy of tese approximations. Ten

More information

WAVELET transforms have become one of the basic tools

WAVELET transforms have become one of the basic tools IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL 13, NO 7, JULY 2004 993 A New Family of Spline-Based Biorthogonal Wavelet Transforms and Their Application to Image Compression Amir Z Averbuch and Valery A Zheludev

More information

Two Modifications of Weight Calculation of the Non-Local Means Denoising Method

Two Modifications of Weight Calculation of the Non-Local Means Denoising Method Engineering, 2013, 5, 522-526 ttp://dx.doi.org/10.4236/eng.2013.510b107 Publised Online October 2013 (ttp://www.scirp.org/journal/eng) Two Modifications of Weigt Calculation of te Non-Local Means Denoising

More information

3.6 Directional Derivatives and the Gradient Vector

3.6 Directional Derivatives and the Gradient Vector 288 CHAPTER 3. FUNCTIONS OF SEVERAL VARIABLES 3.6 Directional Derivatives and te Gradient Vector 3.6.1 Functions of two Variables Directional Derivatives Let us first quickly review, one more time, te

More information

Cubic smoothing spline

Cubic smoothing spline Cubic smooting spline Menu: QCExpert Regression Cubic spline e module Cubic Spline is used to fit any functional regression curve troug data wit one independent variable x and one dependent random variable

More information

Section 2.3: Calculating Limits using the Limit Laws

Section 2.3: Calculating Limits using the Limit Laws Section 2.3: Calculating Limits using te Limit Laws In previous sections, we used graps and numerics to approimate te value of a it if it eists. Te problem wit tis owever is tat it does not always give

More information

More on Functions and Their Graphs

More on Functions and Their Graphs More on Functions and Teir Graps Difference Quotient ( + ) ( ) f a f a is known as te difference quotient and is used exclusively wit functions. Te objective to keep in mind is to factor te appearing in

More information

4.1 Tangent Lines. y 2 y 1 = y 2 y 1

4.1 Tangent Lines. y 2 y 1 = y 2 y 1 41 Tangent Lines Introduction Recall tat te slope of a line tells us ow fast te line rises or falls Given distinct points (x 1, y 1 ) and (x 2, y 2 ), te slope of te line troug tese two points is cange

More information

The Euler and trapezoidal stencils to solve d d x y x = f x, y x

The Euler and trapezoidal stencils to solve d d x y x = f x, y x restart; Te Euler and trapezoidal stencils to solve d d x y x = y x Te purpose of tis workseet is to derive te tree simplest numerical stencils to solve te first order d equation y x d x = y x, and study

More information

MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2

MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2 MATH 5a Spring 2018 READING ASSIGNMENTS FOR CHAPTER 2 Note: Tere will be a very sort online reading quiz (WebWork) on eac reading assignment due one our before class on its due date. Due dates can be found

More information

Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number

Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number Bounding Tree Cover Number and Positive Semidefinite Zero Forcing Number Sofia Burille Mentor: Micael Natanson September 15, 2014 Abstract Given a grap, G, wit a set of vertices, v, and edges, various

More information

Fast Calculation of Thermodynamic Properties of Water and Steam in Process Modelling using Spline Interpolation

Fast Calculation of Thermodynamic Properties of Water and Steam in Process Modelling using Spline Interpolation P R E P R N T CPWS XV Berlin, September 8, 008 Fast Calculation of Termodynamic Properties of Water and Steam in Process Modelling using Spline nterpolation Mattias Kunick a, Hans-Joacim Kretzscmar a,

More information

, 1 1, A complex fraction is a quotient of rational expressions (including their sums) that result

, 1 1, A complex fraction is a quotient of rational expressions (including their sums) that result RT. Complex Fractions Wen working wit algebraic expressions, sometimes we come across needing to simplify expressions like tese: xx 9 xx +, xx + xx + xx, yy xx + xx + +, aa Simplifying Complex Fractions

More information

Grid Adaptation for Functional Outputs: Application to Two-Dimensional Inviscid Flows

Grid Adaptation for Functional Outputs: Application to Two-Dimensional Inviscid Flows Journal of Computational Pysics 176, 40 69 (2002) doi:10.1006/jcp.2001.6967, available online at ttp://www.idealibrary.com on Grid Adaptation for Functional Outputs: Application to Two-Dimensional Inviscid

More information

ANTENNA SPHERICAL COORDINATE SYSTEMS AND THEIR APPLICATION IN COMBINING RESULTS FROM DIFFERENT ANTENNA ORIENTATIONS

ANTENNA SPHERICAL COORDINATE SYSTEMS AND THEIR APPLICATION IN COMBINING RESULTS FROM DIFFERENT ANTENNA ORIENTATIONS NTNN SPHRICL COORDINT SSTMS ND THIR PPLICTION IN COMBINING RSULTS FROM DIFFRNT NTNN ORINTTIONS llen C. Newell, Greg Hindman Nearfield Systems Incorporated 133. 223 rd St. Bldg. 524 Carson, C 9745 US BSTRCT

More information

CE 221 Data Structures and Algorithms

CE 221 Data Structures and Algorithms CE Data Structures and Algoritms Capter 4: Trees (AVL Trees) Text: Read Weiss, 4.4 Izmir University of Economics AVL Trees An AVL (Adelson-Velskii and Landis) tree is a binary searc tree wit a balance

More information

16th European Signal Processing Conference (EUSIPCO 2008), Lausanne, Switzerland, August 25-29, 2008, copyright by EURASIP

16th European Signal Processing Conference (EUSIPCO 2008), Lausanne, Switzerland, August 25-29, 2008, copyright by EURASIP 16t European Signal Processing Conference (EUSIPCO 008), Lausanne, Switzerland, August 5-9, 008, copyrigt by EURASIP ADAPTIVE WINDOW FOR LOCAL POLYNOMIAL REGRESSION FROM NOISY NONUNIFORM SAMPLES A. Sreenivasa

More information

Haar Transform CS 430 Denbigh Starkey

Haar Transform CS 430 Denbigh Starkey Haar Transform CS Denbig Starkey. Background. Computing te transform. Restoring te original image from te transform 7. Producing te transform matrix 8 5. Using Haar for lossless compression 6. Using Haar

More information

An Explicit Formula for Generalized Arithmetic-Geometric Sum

An Explicit Formula for Generalized Arithmetic-Geometric Sum Applied Matematical Sciences, Vol. 9, 2015, no. 114, 5687-5696 HIKARI Ltd, www.m-ikari.com ttp://dx.doi.org/10.12988/ams.2015.57481 An Explicit Formula for Generalized Aritmetic-Geometric Sum Roberto B.

More information

Alternating Direction Implicit Methods for FDTD Using the Dey-Mittra Embedded Boundary Method

Alternating Direction Implicit Methods for FDTD Using the Dey-Mittra Embedded Boundary Method Te Open Plasma Pysics Journal, 2010, 3, 29-35 29 Open Access Alternating Direction Implicit Metods for FDTD Using te Dey-Mittra Embedded Boundary Metod T.M. Austin *, J.R. Cary, D.N. Smite C. Nieter Tec-X

More information

Multi-Stack Boundary Labeling Problems

Multi-Stack Boundary Labeling Problems Multi-Stack Boundary Labeling Problems Micael A. Bekos 1, Micael Kaufmann 2, Katerina Potika 1 Antonios Symvonis 1 1 National Tecnical University of Atens, Scool of Applied Matematical & Pysical Sciences,

More information

4.2 The Derivative. f(x + h) f(x) lim

4.2 The Derivative. f(x + h) f(x) lim 4.2 Te Derivative Introduction In te previous section, it was sown tat if a function f as a nonvertical tangent line at a point (x, f(x)), ten its slope is given by te it f(x + ) f(x). (*) Tis is potentially

More information

Efficient and Low-Complexity Image Coding with the Lifting Scheme and Modified SPIHT

Efficient and Low-Complexity Image Coding with the Lifting Scheme and Modified SPIHT Efficient and Low-Complexity Image Coding with the Lifting Scheme and Modified SPIHT Hong Pan, W.C. Siu, and N.F. Law Abstract In this paper, we propose an efficient and low complexity image coding algorithm

More information

The (, D) and (, N) problems in double-step digraphs with unilateral distance

The (, D) and (, N) problems in double-step digraphs with unilateral distance Electronic Journal of Grap Teory and Applications () (), Te (, D) and (, N) problems in double-step digraps wit unilateral distance C Dalfó, MA Fiol Departament de Matemàtica Aplicada IV Universitat Politècnica

More information

MAPI Computer Vision

MAPI Computer Vision MAPI Computer Vision Multiple View Geometry In tis module we intend to present several tecniques in te domain of te 3D vision Manuel Joao University of Mino Dep Industrial Electronics - Applications -

More information

2.8 The derivative as a function

2.8 The derivative as a function CHAPTER 2. LIMITS 56 2.8 Te derivative as a function Definition. Te derivative of f(x) istefunction f (x) defined as follows f f(x + ) f(x) (x). 0 Note: tis differs from te definition in section 2.7 in

More information

All truths are easy to understand once they are discovered; the point is to discover them. Galileo

All truths are easy to understand once they are discovered; the point is to discover them. Galileo Section 7. olume All truts are easy to understand once tey are discovered; te point is to discover tem. Galileo Te main topic of tis section is volume. You will specifically look at ow to find te volume

More information

AVL Trees Outline and Required Reading: AVL Trees ( 11.2) CSE 2011, Winter 2017 Instructor: N. Vlajic

AVL Trees Outline and Required Reading: AVL Trees ( 11.2) CSE 2011, Winter 2017 Instructor: N. Vlajic 1 AVL Trees Outline and Required Reading: AVL Trees ( 11.2) CSE 2011, Winter 2017 Instructor: N. Vlajic AVL Trees 2 Binary Searc Trees better tan linear dictionaries; owever, te worst case performance

More information

Unsupervised Learning for Hierarchical Clustering Using Statistical Information

Unsupervised Learning for Hierarchical Clustering Using Statistical Information Unsupervised Learning for Hierarcical Clustering Using Statistical Information Masaru Okamoto, Nan Bu, and Tosio Tsuji Department of Artificial Complex System Engineering Hirosima University Kagamiyama

More information

1.4 RATIONAL EXPRESSIONS

1.4 RATIONAL EXPRESSIONS 6 CHAPTER Fundamentals.4 RATIONAL EXPRESSIONS Te Domain of an Algebraic Epression Simplifying Rational Epressions Multiplying and Dividing Rational Epressions Adding and Subtracting Rational Epressions

More information

( ) C 2 ( ) C 1. cllqfkd=rkabo=^uf^i=il^a=^ka=jljbkq=er. KlingnerF=

( ) C 2 ( ) C 1. cllqfkd=rkabo=^uf^i=il^a=^ka=jljbkq=er. KlingnerF= cllqfkd=rkabo=^uf^i=il^a=^ka=jljbkq=er. KlingnerF= A reinfoced concrete footing needs to be designed to support column A3 in te indicated drawing. According to te preliminary structural analysis tat was

More information

Investigating an automated method for the sensitivity analysis of functions

Investigating an automated method for the sensitivity analysis of functions Investigating an automated metod for te sensitivity analysis of functions Sibel EKER s.eker@student.tudelft.nl Jill SLINGER j..slinger@tudelft.nl Delft University of Tecnology 2628 BX, Delft, te Neterlands

More information

Materials: Whiteboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector.

Materials: Whiteboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector. Adam Clinc Lesson: Deriving te Derivative Grade Level: 12 t grade, Calculus I class Materials: Witeboard, TI-Nspire classroom set, quadratic tangents program, and a computer projector. Goals/Objectives:

More information

Proceedings of the 8th WSEAS International Conference on Neural Networks, Vancouver, British Columbia, Canada, June 19-21,

Proceedings of the 8th WSEAS International Conference on Neural Networks, Vancouver, British Columbia, Canada, June 19-21, Proceedings of te 8t WSEAS International Conference on Neural Networks, Vancouver, Britis Columbia, Canada, June 9-2, 2007 3 Neural Network Structures wit Constant Weigts to Implement Dis-Jointly Removed

More information

Computing geodesic paths on manifolds

Computing geodesic paths on manifolds Proc. Natl. Acad. Sci. USA Vol. 95, pp. 8431 8435, July 1998 Applied Matematics Computing geodesic pats on manifolds R. Kimmel* and J. A. Setian Department of Matematics and Lawrence Berkeley National

More information

RECENTLY frames or redundant expansions of signals

RECENTLY frames or redundant expansions of signals 2126 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 54, NO. 6, JUNE 2006 Interpolatory Frames in Signal Space Amir Z. Averbuch, Valery A. Zheludev, Tamir Cohen Abstract We present a new family of frames,

More information

Erasing Haar Coefficients

Erasing Haar Coefficients Recap Haar simple and fast wavelet transform Limitations not smooth enough: blocky How to improve? classical approach: basis functions Lifting: transforms 1 Erasing Haar Coefficients 2 Page 1 Classical

More information

19.2 Surface Area of Prisms and Cylinders

19.2 Surface Area of Prisms and Cylinders Name Class Date 19 Surface Area of Prisms and Cylinders Essential Question: How can you find te surface area of a prism or cylinder? Resource Locker Explore Developing a Surface Area Formula Surface area

More information

Key words: B- Spline filters, filter banks, sub band coding, Pre processing, Image Averaging IJSER

Key words: B- Spline filters, filter banks, sub band coding, Pre processing, Image Averaging IJSER International Journal of Scientific & Engineering Research, Volume 7, Issue 9, September-2016 470 Analyzing Low Bit Rate Image Compression Using Filters and Pre Filtering PNV ABHISHEK 1, U VINOD KUMAR

More information

Minimizing Memory Access By Improving Register Usage Through High-level Transformations

Minimizing Memory Access By Improving Register Usage Through High-level Transformations Minimizing Memory Access By Improving Register Usage Troug Hig-level Transformations San Li Scool of Computer Engineering anyang Tecnological University anyang Avenue, SIGAPORE 639798 Email: p144102711@ntu.edu.sg

More information

H-Adaptive Multiscale Schemes for the Compressible Navier-Stokes Equations Polyhedral Discretization, Data Compression and Mesh Generation

H-Adaptive Multiscale Schemes for the Compressible Navier-Stokes Equations Polyhedral Discretization, Data Compression and Mesh Generation H-Adaptive Multiscale Scemes for te Compressible Navier-Stokes Equations Polyedral Discretization, Data Compression and Mes Generation F. Bramkamp 1, B. Gottsclic-Müller 2, M. Hesse 1, P. Lamby 2, S. Müller

More information

2 The Derivative. 2.0 Introduction to Derivatives. Slopes of Tangent Lines: Graphically

2 The Derivative. 2.0 Introduction to Derivatives. Slopes of Tangent Lines: Graphically 2 Te Derivative Te two previous capters ave laid te foundation for te study of calculus. Tey provided a review of some material you will need and started to empasize te various ways we will view and use

More information

Optimal In-Network Packet Aggregation Policy for Maximum Information Freshness

Optimal In-Network Packet Aggregation Policy for Maximum Information Freshness 1 Optimal In-etwork Packet Aggregation Policy for Maimum Information Fresness Alper Sinan Akyurek, Tajana Simunic Rosing Electrical and Computer Engineering, University of California, San Diego aakyurek@ucsd.edu,

More information

A Novel QC-LDPC Code with Flexible Construction and Low Error Floor

A Novel QC-LDPC Code with Flexible Construction and Low Error Floor A Novel QC-LDPC Code wit Flexile Construction and Low Error Floor Hanxin WANG,2, Saoping CHEN,2,CuitaoZHU,2 and Kaiyou SU Department of Electronics and Information Engineering, Sout-Central University

More information

You Try: A. Dilate the following figure using a scale factor of 2 with center of dilation at the origin.

You Try: A. Dilate the following figure using a scale factor of 2 with center of dilation at the origin. 1 G.SRT.1-Some Tings To Know Dilations affect te size of te pre-image. Te pre-image will enlarge or reduce by te ratio given by te scale factor. A dilation wit a scale factor of 1> x >1enlarges it. A dilation

More information

Tuning MAX MIN Ant System with off-line and on-line methods

Tuning MAX MIN Ant System with off-line and on-line methods Université Libre de Bruxelles Institut de Recerces Interdisciplinaires et de Développements en Intelligence Artificielle Tuning MAX MIN Ant System wit off-line and on-line metods Paola Pellegrini, Tomas

More information

HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS

HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS HASH ALGORITHMS: A DESIGN FOR PARALLEL CALCULATIONS N.G.Bardis Researc Associate Hellenic Ministry of te Interior, Public Administration and Decentralization 8, Dragatsaniou str., Klatmonos S. 0559, Greece

More information

Chapter K. Geometric Optics. Blinn College - Physics Terry Honan

Chapter K. Geometric Optics. Blinn College - Physics Terry Honan Capter K Geometric Optics Blinn College - Pysics 2426 - Terry Honan K. - Properties of Ligt Te Speed of Ligt Te speed of ligt in a vacuum is approximately c > 3.0µ0 8 mês. Because of its most fundamental

More information

Symmetric Tree Replication Protocol for Efficient Distributed Storage System*

Symmetric Tree Replication Protocol for Efficient Distributed Storage System* ymmetric Tree Replication Protocol for Efficient Distributed torage ystem* ung Cune Coi 1, Hee Yong Youn 1, and Joong up Coi 2 1 cool of Information and Communications Engineering ungkyunkwan University

More information

ICES REPORT Isogeometric Analysis of Boundary Integral Equations

ICES REPORT Isogeometric Analysis of Boundary Integral Equations ICES REPORT 5-2 April 205 Isogeometric Analysis of Boundary Integral Equations by Mattias Taus, Gregory J. Rodin and Tomas J. R. Huges Te Institute for Computational Engineering and Sciences Te University

More information

An Algorithm for Loopless Deflection in Photonic Packet-Switched Networks

An Algorithm for Loopless Deflection in Photonic Packet-Switched Networks An Algoritm for Loopless Deflection in Potonic Packet-Switced Networks Jason P. Jue Center for Advanced Telecommunications Systems and Services Te University of Texas at Dallas Ricardson, TX 75083-0688

More information

Tilings of rectangles with T-tetrominoes

Tilings of rectangles with T-tetrominoes Tilings of rectangles wit T-tetrominoes Micael Korn and Igor Pak Department of Matematics Massacusetts Institute of Tecnology Cambridge, MA, 2139 mikekorn@mit.edu, pak@mat.mit.edu August 26, 23 Abstract

More information

THE POSSIBILITY OF ESTIMATING THE VOLUME OF A SQUARE FRUSTRUM USING THE KNOWN VOLUME OF A CONICAL FRUSTRUM

THE POSSIBILITY OF ESTIMATING THE VOLUME OF A SQUARE FRUSTRUM USING THE KNOWN VOLUME OF A CONICAL FRUSTRUM THE POSSIBILITY OF ESTIMATING THE VOLUME OF A SQUARE FRUSTRUM USING THE KNOWN VOLUME OF A CONICAL FRUSTRUM SAMUEL OLU OLAGUNJU Adeyemi College of Education NIGERIA Email: lagsam04@aceondo.edu.ng ABSTRACT

More information

3. Lifting Scheme of Wavelet Transform

3. Lifting Scheme of Wavelet Transform 3. Lifting Scheme of Wavelet Transform 3. Introduction The Wim Sweldens 76 developed the lifting scheme for the construction of biorthogonal wavelets. The main feature of the lifting scheme is that all

More information

A C 2 Four-Point Subdivision Scheme with Fourth Order Accuracy and its Extensions

A C 2 Four-Point Subdivision Scheme with Fourth Order Accuracy and its Extensions A C 2 Four-Point Subdivision Scheme with Fourth Order Accuracy and its Extensions Nira Dyn School of Mathematical Sciences Tel Aviv University Michael S. Floater Department of Informatics University of

More information

Inference Based on SVARs Identified with Sign and Zero Restrictions: Theory and Applications

Inference Based on SVARs Identified with Sign and Zero Restrictions: Theory and Applications FEDERAL RESERVE BANK of ATLANTA WORKING PAPER SERIES Inference Based on SVARs Identified wit Sign and Zero Restrictions: Teory and Applications Jonas E. Arias, Juan F. Rubio-Ramírez, and Daniel F. Waggoner

More information

Non-Interferometric Testing

Non-Interferometric Testing NonInterferometric Testing.nb Optics 513 - James C. Wyant 1 Non-Interferometric Testing Introduction In tese notes four non-interferometric tests are described: (1) te Sack-Hartmann test, (2) te Foucault

More information

Zernike vs. Zonal Matrix Iterative Wavefront Reconstructor. Sophia I. Panagopoulou, PhD. University of Crete Medical School Dept.

Zernike vs. Zonal Matrix Iterative Wavefront Reconstructor. Sophia I. Panagopoulou, PhD. University of Crete Medical School Dept. Zernie vs. Zonal Matrix terative Wavefront Reconstructor opia. Panagopoulou PD University of Crete Medical cool Dept. of Optalmology Daniel R. Neal PD Wavefront ciences nc. 480 Central.E. Albuquerque NM

More information

VHDL Implementation of Multiplierless, High Performance DWT Filter Bank

VHDL Implementation of Multiplierless, High Performance DWT Filter Bank VHDL Implementation of Multiplierless, High Performance DWT Filter Bank Mr. M.M. Aswale 1, Prof. Ms. R.B Patil 2,Member ISTE Abstract The JPEG 2000 image coding standard employs the biorthogonal 9/7 wavelet

More information

Intra- and Inter-Session Network Coding in Wireless Networks

Intra- and Inter-Session Network Coding in Wireless Networks Intra- and Inter-Session Network Coding in Wireless Networks Hulya Seferoglu, Member, IEEE, Atina Markopoulou, Member, IEEE, K K Ramakrisnan, Fellow, IEEE arxiv:857v [csni] 3 Feb Abstract In tis paper,

More information

13.5 DIRECTIONAL DERIVATIVES and the GRADIENT VECTOR

13.5 DIRECTIONAL DERIVATIVES and the GRADIENT VECTOR 13.5 Directional Derivatives and te Gradient Vector Contemporary Calculus 1 13.5 DIRECTIONAL DERIVATIVES and te GRADIENT VECTOR Directional Derivatives In Section 13.3 te partial derivatives f x and f

More information

Overcomplete Steerable Pyramid Filters and Rotation Invariance

Overcomplete Steerable Pyramid Filters and Rotation Invariance vercomplete Steerable Pyramid Filters and Rotation Invariance H. Greenspan, S. Belongie R. Goodman and P. Perona S. Raksit and C. H. Anderson Department of Electrical Engineering Department of Anatomy

More information

Obstacle Avoiding Real-Time Trajectory Generation and Control of Omnidirectional Vehicles

Obstacle Avoiding Real-Time Trajectory Generation and Control of Omnidirectional Vehicles 2009 American Control Conference Hyatt Regency Riverfront, St. Louis, MO, USA June 10-12, 2009 FrC10.6 Obstacle Avoiding Real-Time Trajectory Generation and Control of Omnidirectional Veicles Ji-wung Coi,

More information

Mean Shifting Gradient Vector Flow: An Improved External Force Field for Active Surfaces in Widefield Microscopy.

Mean Shifting Gradient Vector Flow: An Improved External Force Field for Active Surfaces in Widefield Microscopy. Mean Sifting Gradient Vector Flow: An Improved External Force Field for Active Surfaces in Widefield Microscopy. Margret Keuper Cair of Pattern Recognition and Image Processing Computer Science Department

More information

CESILA: Communication Circle External Square Intersection-Based WSN Localization Algorithm

CESILA: Communication Circle External Square Intersection-Based WSN Localization Algorithm Sensors & Transducers 2013 by IFSA ttp://www.sensorsportal.com CESILA: Communication Circle External Square Intersection-Based WSN Localization Algoritm Sun Hongyu, Fang Ziyi, Qu Guannan College of Computer

More information

Redundancy Awareness in SQL Queries

Redundancy Awareness in SQL Queries Redundancy Awareness in QL Queries Bin ao and Antonio Badia omputer Engineering and omputer cience Department University of Louisville bin.cao,abadia @louisville.edu Abstract In tis paper, we study QL

More information

Harmonic Spline Series Representation of Scaling Functions

Harmonic Spline Series Representation of Scaling Functions Harmonic Spline Series Representation of Scaling Functions Thierry Blu and Michael Unser Biomedical Imaging Group, STI/BIO-E, BM 4.34 Swiss Federal Institute of Technology, Lausanne CH-5 Lausanne-EPFL,

More information

Section 1.2 The Slope of a Tangent

Section 1.2 The Slope of a Tangent Section 1.2 Te Slope of a Tangent You are familiar wit te concept of a tangent to a curve. Wat geometric interpretation can be given to a tangent to te grap of a function at a point? A tangent is te straigt

More information

Data Structures and Programming Spring 2014, Midterm Exam.

Data Structures and Programming Spring 2014, Midterm Exam. Data Structures and Programming Spring 2014, Midterm Exam. 1. (10 pts) Order te following functions 2.2 n, log(n 10 ), 2 2012, 25n log(n), 1.1 n, 2n 5.5, 4 log(n), 2 10, n 1.02, 5n 5, 76n, 8n 5 + 5n 2

More information

On the use of FHT, its modification for practical applications and the structure of Hough image

On the use of FHT, its modification for practical applications and the structure of Hough image On te use of FHT, its modification for practical applications and te structure of Houg image M. Aliev 1,3, E.I. Ersov, D.P. Nikolaev,3 1 Federal Researc Center Computer Science and Control of Russian Academy

More information

Implementation of Integral based Digital Curvature Estimators in DGtal

Implementation of Integral based Digital Curvature Estimators in DGtal Implementation of Integral based Digital Curvature Estimators in DGtal David Coeurjolly 1, Jacques-Olivier Lacaud 2, Jérémy Levallois 1,2 1 Université de Lyon, CNRS INSA-Lyon, LIRIS, UMR5205, F-69621,

More information

SIGNAL COMPRESSION. 9. Lossy image compression: SPIHT and S+P

SIGNAL COMPRESSION. 9. Lossy image compression: SPIHT and S+P SIGNAL COMPRESSION 9. Lossy image compression: SPIHT and S+P 9.1 SPIHT embedded coder 9.2 The reversible multiresolution transform S+P 9.3 Error resilience in embedded coding 178 9.1 Embedded Tree-Based

More information

Density Estimation Over Data Stream

Density Estimation Over Data Stream Density Estimation Over Data Stream Aoying Zou Dept. of Computer Science, Fudan University 22 Handan Rd. Sangai, 2433, P.R. Cina ayzou@fudan.edu.cn Ziyuan Cai Dept. of Computer Science, Fudan University

More information

THANK YOU FOR YOUR PURCHASE!

THANK YOU FOR YOUR PURCHASE! THANK YOU FOR YOUR PURCHASE! Te resources included in tis purcase were designed and created by me. I ope tat you find tis resource elpful in your classroom. Please feel free to contact me wit any questions

More information

Inference Based on SVARs Identified with Sign and Zero Restrictions: Theory and Applications

Inference Based on SVARs Identified with Sign and Zero Restrictions: Theory and Applications Inference Based on SVARs Identified wit Sign and Zero Restrictions: Teory and Applications Jonas E. Arias Federal Reserve Board Juan F. Rubio-Ramírez Emory University, BBVA Researc, and Federal Reserve

More information

Distributed and Optimal Rate Allocation in Application-Layer Multicast

Distributed and Optimal Rate Allocation in Application-Layer Multicast Distributed and Optimal Rate Allocation in Application-Layer Multicast Jinyao Yan, Martin May, Bernard Plattner, Wolfgang Mülbauer Computer Engineering and Networks Laboratory, ETH Zuric, CH-8092, Switzerland

More information

Coarticulation: An Approach for Generating Concurrent Plans in Markov Decision Processes

Coarticulation: An Approach for Generating Concurrent Plans in Markov Decision Processes Coarticulation: An Approac for Generating Concurrent Plans in Markov Decision Processes Kasayar Roanimanes kas@cs.umass.edu Sridar Maadevan maadeva@cs.umass.edu Department of Computer Science, University

More information

Pseudo-Developing Maps for Ideal Triangulations I: Essential Edges and Generalised Hyperbolic Gluing Equations

Pseudo-Developing Maps for Ideal Triangulations I: Essential Edges and Generalised Hyperbolic Gluing Equations Pseudo-Developing Maps for Ideal Triangulations I: Essential Edges and Generalised Hyperbolic Gluing Equations Henry Segerman and Stepan Tillmann To Bus Jaco on te occasion of is seventiet birtday arxiv:1107.1030v1

More information

Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks

Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks Utilizing Call Admission Control to Derive Optimal Pricing of Multiple Service Classes in Wireless Cellular Networks Okan Yilmaz and Ing-Ray Cen Computer Science Department Virginia Tec {oyilmaz, ircen}@vt.edu

More information

On the Use of Radio Resource Tests in Wireless ad hoc Networks

On the Use of Radio Resource Tests in Wireless ad hoc Networks Tecnical Report RT/29/2009 On te Use of Radio Resource Tests in Wireless ad oc Networks Diogo Mónica diogo.monica@gsd.inesc-id.pt João Leitão jleitao@gsd.inesc-id.pt Luis Rodrigues ler@ist.utl.pt Carlos

More information

An Anchor Chain Scheme for IP Mobility Management

An Anchor Chain Scheme for IP Mobility Management An Ancor Cain Sceme for IP Mobility Management Yigal Bejerano and Israel Cidon Department of Electrical Engineering Tecnion - Israel Institute of Tecnology Haifa 32000, Israel E-mail: bej@tx.tecnion.ac.il.

More information

VOLUMES. The volume of a cylinder is determined by multiplying the cross sectional area by the height. r h V. a) 10 mm 25 mm.

VOLUMES. The volume of a cylinder is determined by multiplying the cross sectional area by the height. r h V. a) 10 mm 25 mm. OLUME OF A CYLINDER OLUMES Te volume of a cylinder is determined by multiplying te cross sectional area by te eigt. r Were: = volume r = radius = eigt Exercise 1 Complete te table ( =.14) r a) 10 mm 5

More information

A C 2 Four-Point Subdivision Scheme with Fourth Order Accuracy and its Extensions

A C 2 Four-Point Subdivision Scheme with Fourth Order Accuracy and its Extensions A C 2 Four-Point Subdivision Scheme with Fourth Order Accuracy and its Extensions Nira Dyn Michael S. Floater Kai Hormann Abstract. We present a new four-point subdivision scheme that generates C 2 curves.

More information

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Spring

Hash-Based Indexes. Chapter 11. Comp 521 Files and Databases Spring Has-Based Indexes Capter 11 Comp 521 Files and Databases Spring 2010 1 Introduction As for any index, 3 alternatives for data entries k*: Data record wit key value k

More information

Asynchronous Power Flow on Graphic Processing Units

Asynchronous Power Flow on Graphic Processing Units 1 Asyncronous Power Flow on Grapic Processing Units Manuel Marin, Student Member, IEEE, David Defour, and Federico Milano, Senior Member, IEEE Abstract Asyncronous iterations can be used to implement fixed-point

More information

Multi-Objective Particle Swarm Optimizers: A Survey of the State-of-the-Art

Multi-Objective Particle Swarm Optimizers: A Survey of the State-of-the-Art Multi-Objective Particle Swarm Optimizers: A Survey of te State-of-te-Art Margarita Reyes-Sierra and Carlos A. Coello Coello CINVESTAV-IPN (Evolutionary Computation Group) Electrical Engineering Department,

More information

CHAPTER 7: TRANSCENDENTAL FUNCTIONS

CHAPTER 7: TRANSCENDENTAL FUNCTIONS 7.0 Introduction and One to one Functions Contemporary Calculus 1 CHAPTER 7: TRANSCENDENTAL FUNCTIONS Introduction In te previous capters we saw ow to calculate and use te derivatives and integrals of

More information

Image Registration via Particle Movement

Image Registration via Particle Movement Image Registration via Particle Movement Zao Yi and Justin Wan Abstract Toug fluid model offers a good approac to nonrigid registration wit large deformations, it suffers from te blurring artifacts introduced

More information

A UPnP-based Decentralized Service Discovery Improved Algorithm

A UPnP-based Decentralized Service Discovery Improved Algorithm Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol.1, No.1, Marc 2013, pp. 21~26 ISSN: 2089-3272 21 A UPnP-based Decentralized Service Discovery Improved Algoritm Yu Si-cai*, Wu Yan-zi,

More information

12.2 Techniques for Evaluating Limits

12.2 Techniques for Evaluating Limits 335_qd /4/5 :5 PM Page 863 Section Tecniques for Evaluating Limits 863 Tecniques for Evaluating Limits Wat ou sould learn Use te dividing out tecnique to evaluate its of functions Use te rationalizing

More information

12.2 TECHNIQUES FOR EVALUATING LIMITS

12.2 TECHNIQUES FOR EVALUATING LIMITS Section Tecniques for Evaluating Limits 86 TECHNIQUES FOR EVALUATING LIMITS Wat ou sould learn Use te dividing out tecnique to evaluate its of functions Use te rationalizing tecnique to evaluate its of

More information

Energy efficient temporal load aware resource allocation in cloud computing datacenters

Energy efficient temporal load aware resource allocation in cloud computing datacenters Vakilinia Journal of Cloud Computing: Advances, Systems and Applications (2018) 7:2 DOI 10.1186/s13677-017-0103-2 Journal of Cloud Computing: Advances, Systems and Applications RESEARCH Energy efficient

More information

Cooperation in Wireless Ad Hoc Networks

Cooperation in Wireless Ad Hoc Networks Cooperation in Wireless Ad Hoc Networks Vikram Srinivasan, Pavan Nuggealli, Carla F. Ciasserini, Rames R. Rao Department of Electrical and Computer Engineering, University of California at San Diego email:

More information

Inference Based on SVARs Identified with Sign and Zero Restrictions: Theory and Applications

Inference Based on SVARs Identified with Sign and Zero Restrictions: Theory and Applications Inference Based on SVARs Identified wit Sign and Zero Restrictions: Teory and Applications Jonas E. Arias Federal Reserve Board Juan F. Rubio-Ramírez Emory University, BBVA Researc, and Federal Reserve

More information

Design of Orthogonal Graph Wavelet Filter Banks

Design of Orthogonal Graph Wavelet Filter Banks Design of Orthogonal Graph Wavelet Filter Banks Xi ZHANG Department of Communication Engineering and Informatics The University of Electro-Communications Chofu-shi, Tokyo, 182-8585 JAPAN E-mail: zhangxi@uec.ac.jp

More information